Websites for real estate are getting crawled by Perplexity through building data-heavy pages, which are answer-based and use schema RealEstateListing and RealEstateAgent (which PerplexityBot can then crawl) and updating these pages frequently. For each query, Perplexity crawls ten to twenty web-pages and references only three to four of those web-pages; the preference is for newer and more descriptive content which has been specifically referenced versus a standard listing page.
What makes real estate websites so non-existent in AI search results? Real estate is the biggest asset class in America, however it ranks dead last among other asset classes in terms of how visible they are in AI search. A recent report from Haute Living and 5W Public Relations states that based upon a report titled “2026 Luxury Real Estate AI Discovery,” real estate accounts for 0.14% of the time that an AI overview is triggered via Google search. Health represents 13%, finance 4.2% and retail 2.1%. What individuals spend the greatest amount of money on during their lifetime, barely exists within the AI panel.
This is relevant today and not tomorrow. Buyers have already made the move. According to a study completed by FlyDragon titled “State of AI Search in Real Estate” that was based upon tracking eight-point-two-million real estate queries across 192 different US cities, sixty-one point-three percent of all buyer side searches begin inside an AI interface instead of a search box. Based upon a subsequent survey conducted by the same firm of 4,180 home-buyers, sixty-seven percent reported using an AI tool as their primary source of research prior to even contacting an agent. This number increased eighteen months later to seventy percent. That’s not a trend; that’s a migration. Additionally, very few are available to meet them. Research compiled by Atlas Unchained indicates that only eight-point-four percent of active U.S. real estate agents appear in any AI generated answer to a high intent query within their own market; whereas the top one percent of agents account for forty-seven percent of all citations generated by AI tools. Therefore, the names of the top one percent receive read aloud by ChatGPT, Gemini and Perplexity to the exact consumers who every other agent would fight over to fill out a portal lead form.
The portals know this. Zillow became the first real estate application inside ChatGPT in October 2025; Redfin became the second real estate application inside ChatGPT in February 2026; Realtor.com released its own ChatGPT application in March 2026. Each major portal has now become a conversational interface; and each major AI assistant has now become a conversational interface related to real estate.
I have observed numerous changes in consumer search behavior over many years and I believe I am able to identify the type of behavior exhibited here. This behavior exhibits characteristics identical to early Google Local Search in the mid-2000s where business owners who had completed their profile forms, had provided accurate and current information about their company, controlled the future decade. The difference is that the timeframe to take advantage of the opportunity is much smaller. Additionally, the size of the authoritative corpus treated by AI is significantly thinner than any previously established. As such, thin corpus combined with rising demand creates an extraordinary opportunity for a website associated with real estate. The question remains whether your property pages were designed to be extracted as part of the answer or designed to be scrolled through. Most were developed to be scrolled through.
What are some of the ways property pages get selected to be cited when using Perplexity? Unlike how a search engine would provide you with links to potential answers; Perplexity uses a form of what they call Retrieval Augmented Generation (RAG). When you enter a question into Perplexity, it searches for around 10-20 possible candidate pages. After it finds those candidate pages, it uses a ranking system to select approximately three-to-four of those candidate pages to include in the final response. As explained by Prominara’s Source Selection Breakdown. The RAG process allows Perplexity to pull data that is live at the time of the request and also to attribute every reference to its original source. The difference between getting selected for the candidates pool and getting referenced in the final response is where many property pages fail.
There are several factors that determine whether you fall into either category. First and foremost is freshness. According to WP SEO AI’s analysis there is a clear correlation between citations and fresh content. For example content published or updated within the past thirty days receives a statistically significant increase in citations. The time frame compresses even further for topics that move quickly. Therefore a neighborhood page that was last updated in 2024 may be competing against a new page created this week, and is likely going to lose. Second is answer first format. This too has been shown to be the largest influencer on selection, as repeated studies show. Write your answer in the first sentence prior to explaining the story behind it. Third is factual specificity. Specifically in terms of authority, Perplexity uses this as an indicator of authority.
Therefore a page that contains such things as a median price for homes, an average days on market, and a date of sale would receive preference over a page that simply states “this area is highly sought after.”
I admit I have a prejudice toward listing sites that begin their listing description with something like “welcome to your dream home.” I feel the same way as Perplexity. Perplexity cannot retrieve information from emotion, so if another competitor provides you with numbers instead of feelings, you’re likely to find yourself being ignored by Perplexity.
Crawlability is another factor, but a harder one. If PerplexityBot is unable to crawl your website due to being denied access via the inclusion of “robots.txt” in your website’s code, your website will never be included in the list of websites that Perplexity considers during its initial search phase. Corroborating evidence also plays a role in determining whether or not Perplexity will use your content. Since Perplexity weighs references made by multiple separate brands very high, it is beneficial for you to be mentioned by others, rather than solely mentioning yourself. A study done by Semrush of 230,000 + prompts used by Perplexity found that the most commonly cited domains were Reddit, LinkedIn and Wikipedia. In addition a study completed by ZipTie regarding pipelines for Perplexity stated that compared to Google’s 881, Perplexity cites 1430 unique news sources. Therefore Perplexity rewards being discussed versus simply discussing yourself.
While it may seem discouraging, AI citation has decoupled from the raw rankings. A recent GEO study conducted by Starmorph concluded that by February 2026, only 38 percent of the URLs that were cited by AI ranked in the Top Ten Organic results on Google. Furthermore 31% came from positions 11-100 and 31% came from outside position 100. Therefore you don’t necessarily need to rank above Zillow in order for your property page to be cited. All you need to do is present the clearest, freshest, most easily-extractable answer to a particular question. That is a battle that a focused local website can actually compete with.
When we talk about what real estate websites should be doing to be referenced in an artificial intelligence (AI) system we have to think about the page itself before thinking about how it develops its own reputation.
First, put the final result of the article on the first paragraph of every individual property page, neighborhood guide and blog entry.
Perplexity picks up on the first solid fact it can assign to the item being searched, therefore if the answer is hidden inside the body of the writing – it may go unnoticed.
Secondly, create articles based upon data rather than adjectives. For example, a monthly entry called “Median price and days on market for [neighborhood] through June 2026,” with a simple table, will likely receive more references than 10 ‘lifestyle’ entries.
Thirdly, construct your H2 and H3 headings like actual phrases consumers would enter into a search engine when searching for information about their home purchase in 2026.
For instance, “what first-time home buyers should know about [neighborhood] in 2026.” Then the information below that would explain the answer to this question.
Lastly, markup your webpages. Each property page needs real estate listing schema including listing type, price, full postal address, geo coordinates, floor size and property category. In addition to real estate agent or person schema for the agent. Make sure that both the price and availability are the same on your schema as well as your webpage. Validate your schema using Google’s Rich Results Test.
Stale schema is worse than no schema at all; you’re giving the AI incorrect information.
Sixth, obtain at least One reference to a credible third party reference source. Corroborating sources move you from listed to trusted.
Do not do the opposite of everything i just said. Don’t prevent AI crawlers from crawling your website.
Don’t create “contact us for quote” pages that don’t contain any useful/extractable information. Vague/merchandising-based copy is ignored by machines.
Don’t allow schema to become stale.
And most importantly, realize that quantity doesn’t always equal quality. It’s the brands receiving cites aren’t creating more content, they are creating pages machines can find, validate and utilize.
There is a huge pay-off to creating pages that can be validated and reused by AI. According to starmorph, AI-referral traffic increased by 527% year over year and converts at 4.4x to 5x greater rates than traditional organic search referrals. Additionally, according to FlyDragon, agents/brokerages cited by AI receive 35%-91% more clicks per impression and have approximately a 4.2x increase in 90-day close rates.
One caveat will keep you accountable: columbia journalism review reported a 37% error rate in perplexity’s answers. This is precisely why developing clean/simple/corroborated pages is essential.
Those real estate agents and brokers losing ground in the next two years won’t lose to AI replacing them – they will lose simply because they were invisible at the precise moment that buyer was searching for their information. Create listings and web content that AI systems can reference – and you’ll stop competing for the lead position – you’ll be the answer to it.
Frequently Asked Questions
Question: Can a small real estate site get cited by Perplexity, or is it only big portals?
Answer: Small sites can get cited. AI citation has decoupled from raw rankings, with only 38% of AI-cited URLs ranking in Google’s top 10 as of February 2026. A focused local site that answers a specific question clearly and freshly can be cited alongside or instead of a major portal.
Question: How long does it take to start showing up in AI search?
Answer: Because Perplexity uses live retrieval, optimized content can begin appearing within roughly 1 to 4 weeks. Most professionals see early citation signals within 30 to 90 days of publishing structured, FAQ-rich content, assuming the page is crawlable and indexed.
Question: What schema should property listing pages use?
Answer: Use RealEstateListing schema on each property page, with listing type, price, full address, geo coordinates, floor size, and property category. Layer RealEstateAgent or Person schema for agents, then validate with Google’s Rich Results Test.
Question: Does blocking AI crawlers protect my listings?
Answer: No, it removes you from the answer entirely. If PerplexityBot, GPTBot, or ClaudeBot is blocked in robots.txt, your pages are excluded from retrieval, so they can never be cited.
Question: Is traditional SEO still relevant for AI search?
Answer: Yes. SEO is the foundation. Crawlability, clear structure, and authority decide whether your pages are eligible to be retrieved. AI search adds new signals on top, but it does not replace the fundamentals.
Are you ready to develop your own listings so that they are citable by AI searches? Contact Ongito today